Human epicardial adipose tissue: A review

Harold S. Sacks, MD,a and John N. Fain, PhDb Memphis, TN

We discuss the anatomy, physiology, and pathophysiology of epicardial adipose tissue and its relationship to coronary
atherosclerosis. Epicardial fat stores triglyceride to supply free fatty acids for myocardial energy production and produces
adipokines. Itshares a common embryological origin with mesenteric and omental fat. Likevisceral abdominal fat, epicardial
fat thickness, measured by echocardiography, is increased in obesity. Epicardial fat could influence coronary atherogenesis
and myocardial function because there is no fibrous fascial layer to impede diffusion of free fatty acids and adipokines
between it and the underlying vessel wall as well as the myocardium. Segments of coronary arteries lacking epicardial fat or
separated from it by a bridge of myocardial tissue are protected against the development of atherosclerosis in those
segments. However, when epicardial fat is totally absent in congenital generalized lipodystrophy, coronary atherosclerosis
can stilloccur. Macrophages are more numerous and densely packed in the periadventitial fat of human atherosclerotic
coronary arteries with lipid cores than in that of fibrocalcific or nonatherosclerotic coronary arteries. In obese patients with
multiple cardiovascular risk factors, epicardial fat around atheromatous coronaries secretes several proinflammatory
cytokines and is infiltrated by macrophages, lymphocytes, and basophils. Epicardial adipokine expression in obesity without
coronary atherosclerosis has not been determined. In nonobese patients, epicardial fat around atheromatous coronary
arteries expresses proinflammatory cytokines but produces either less adiponectin, a vasoprotective adipokine, than fat
around non atheromatous coronaries or a similar amount compared with thoracic subcutaneous fat. Further studies should be
done to test the hypothesis that adipokines produced by and released from human epicardial adipose tissue might contribute
locally to the pathogenesis of coronary atherosclerosis. (Am Heart J 2007; 153:907-17.)

Human epicardial adipose tissue (EA1) is a visceral
thoracic fat because of its apposition to the heart, to a
hollow muscular organ, or to the viscus. It has not been
studied as thoroughly I as visceral abdominal adipose
tissue 0' A1) and subcutaneous abdominal adipose tissue
(SCA1).2 Like other white adipose tissue IOci,3-6EAT
might function as a lipid-storing depot, as an endocrine
organ secreting hormones, and as an inflammatory tissue
secreting cytokines and chemokines. Under these con-
ditions, its proximity to the adventitia of the coronary
arteries (Figure 1) and the underlying myocardium
suggests the possibility that it could playa role in the
pathogenesis of coronary atherosclerosis (CAD), itself a
chronic inflammatory disease, I and cardiomyopathy
(CMO). The obesity epidemic in children 7 and adults has
drawn attention to VAT and the metabolic syndromeR as
From the "Divisionof Endocrinologyand Metabolism.Universityof Tennessee,and
Baptist Hospital Heart Insritute,Memphis, TN, and bDepartment of Molecular Sciences,
College of Medicine, University of Tennessee Health Science Center, Memphis. TN.
Disclosure: Harold S. Socks is a member of the Speakers Bureau and has received
honoraria from Takeda Pharmaceuticals and Merck Pharmaceuticals.
Submiffed October 19, 2006; accepted March 13, 2007.
Reprint requests: Harold S. Socks, MD, 6027 Walnut Grove Road, suite 307, Memphis,
TN 38120.
E-mail: hsacks@hotmail.com
0002.B703/$ . see front maffer
<!::>
2007, Mosby, Inc. All rights reserved.
doi: 10. 10 16/j.ahj-2007.03_019
risk factors for cardiovascular disease (CVD) and type 2
diabetes mellitus (DM)9.1Oand poses whether obesity
per se could affect EAT and its adipokine content.
We discuss the anatomy and physiology of human
EAT, the pathophysiology of white adipose tissue in
obesity compared to the nonobese state, the patho-
physiology of EAT, and the putative role of EAT in the
pathogenesis of CAD and CMO.
Anatomy and physiology of EAT
The epicardium or visceral layer of the pericardium is a
population of mesothelial cells that migrate onto the
surface of the heart from the area of the septum
transversum (the embryological source of the dia-
phragm). Epicardial, mesenteric, and omental fat all share
the same origin from the splanchnopleuric mesoderm
associated with the gut. 11 In the normal adult, epicardial
fat is concentrated in the atrioventricular (AV) and
interventricular (IV) grooves and along the major
branches of the coronary arteries, and, to a lesser extent,
around the atria, over the free wall of the right ventricle
(RV) and over the apex of the left ventricle (LV). 12,13
Pericardial fat (pericardial adipose tissue [PAT]) is defined
as epicardial fat in all these possible locations plus
paracardial fat.14 Paracardial fat is situated on the external
surface of the parietal pericardium within the mediasti-
num and has alternatively been termed mediastinal fat.15
908 Sacks and Fain
Figure 1
Histology of the right coronary artery and periadventitial epicardial
fat. This high power (x 100 magni~cation) hematoxylin and eosin
stain of a transverse autopsy section of the right coronary artery from
a patient with hypertensive heart disease shows the layers of the
artery wall, the tissue structures in epicardial fat, and the close
contact of epicardial adipocytes with the adventitia.
Paracardial fat originates from the primitive thoracic
mesenchyme, which splits to form the parietal (fibrous)
pericardium and the outer thoracic wall. 16Epicardial
adipose tissue is supplied by branches of the coronary
arteries, whereas paracardial fat is supplied from different
sources including the pericardiacophrenic artery, a
branch of the internal mammary. 17
Lipolysis and lipogenesis have not been measured
directly in human epicardial fat. Based on approximately
2-fold higher rates of lipolysis and lipogenesis in guinea-
pig epicardial fat than other fat depots, Marchington
et a118,19proposed that EATserves to capture and store
intravascular free fatty acid (FFA) to protect cardiomyo-
cytes from exposure to excessive coronary arterial FFA
concentrations during increased energy intake and, at
other times, to release FFA as an immediate ATP source
for the myocardium during periods of need. Epicardial fat
and the myocardium are contiguous. Islands of mature
adipocytes are more frequent within the subepicardial
myocardium of the RV than the LV13and may act as more
readily available, direct sources of FFA for cardiomyo-
cytes. The thickness of the wall of the right atrium is
about 2 mm; the left atrium, 3 to 5 mm; the RV, 3 to 5 mm;
and the LV, 13 to 15 mm.20 Possibly, FFAs could diffuse
bidirectionally in interstitial fluid across concentration
gradients from epicardial fat into the atrial and RV walls
where EATpredominates and vice versa, but this process
in the LVwall can be questioned because the diffusion
distance is much longer.
American Hearl Journai
June 2007
In normal humans, systemic')1 fat stores are the principal
source ofFFAs for the heart.- The myocardium extracts
and metabolizes FFAsfrom coronary arterial blood. Free
fatty acid kinetic studies show that under normal basal
conditions, endogenous FFAs are released into the
coronary veins and then into the coronary venous
sinus. 21.22The source for this FFArelease is thought to be
EATlipolysis, 22since other possibilities such as hydrolysis
of intracardiomyocyte triglyceride or hydrolysis of circu-
lating very-Iow-density-lipoprotein-triglyceride in coro-
nary blood21 seem unlikely. The reason for FFAefflux into
coronary venous blood is unclear. It might represent an
"overflow" of FFAs not used by the myocardium.
Alternatively, it might be a direct source of FFAs for the
pulmonary arterial circulation, since vasoactive prosta-
noids are generated by the pulmonary arterial endothe-
lium from FFAprecursors.23 The fact that coronary sinus
FFAappearance accounts for a minor fraction of systemic
FFAflUX22supports the hypothesis that EATfunctions as a
local myocardium-specific triglyceride depot. Epicardial
adipose tissue might secrete vasoactive products that
regulate coronary arterial tone. For example, adipocyte-
derived relaxing factor, a protein recently isolated from
normal rodent aortic and mesenteric arterial periadven-
titial fat,24 stimulates arterial vasodilation independently
of nitric oxide by diffusing into the media of the coronary
wall, normally 0.55 to 1.0 mm thick 25It is different from
leptin24 and adiponectin.26
Quantitation of EAT
Autopsy
Corradi et al27 dissected epicardial fat from the
underlying myocardium in a series of 117 patients and
found that it accounted for approximately 15% (mean,
54 I 23 g [ISDJ) of a normal heart weight (365 I 49 g).
They also found a direct correlation between LV and RV
mass and corresponding epicardial fat mass. In a later
study, the same authors confirmed the direct correlation
(r = 0.755, P = .01) between EAT mass and myocardial
ventricular mass measured by echocardiography in
60 subjects with no known cardiac disease.28 In an
unselected group of 200 patients dying from a variety of
diseases including carcinoma and arteriosclerosis studied
by Schjebal,29 epicardial fat thickness over the RV wall
varied from zero along the fat.free diaphragmatic region
to 13.6 mm along the sharp ventrolateral edge close to the
base, the maximal point of thickness. In that report, EAT
thickness correlated directly with subcutaneous fat
thickness and was 1.65-fold greater in women in each of
these locations. Duflou et al30measured EATthickness in
3 selected age-matched groups of subjects: group 1- 22
massively obese adults (mean weight, 175 I 68 kg; body
mass index [BMI], 57 I 12.8 kg/m2 [I SD], currently
classified as morbid obesity) who died suddenly;
group 2-6 massively obese adults (weight, 131 I 25 kg;
American Hearl journal
Volume 153, Number 6
Table I. Correlations between epicardial and pericardial versus visceral abdominal and SCATs
Study (n) 8MI (kg/m2) Radiological
Italian
Men and women (72) 34.0 :t 14.5 (SD)
Italian
Women lean (15) 22.6 :t l.7(SD)
Women obese (27) 43.5 :t 4.8
Italian
Men (23) 27.7 :t 0.6(SEM)
American
Men and women (80) 31.9:t 7.3(SD)
Japanese
Men nonobese (181) 22.7 i 2.0(SD)
Men obese (64) 27.6 :t 2.3(SD)
NS, No statistically significant correlation.
'Measured using MRI.
tMeasured using CT.
!Measured as mediastinol (paracardiol) adipose tissue.
§Measured as paracardial plus epicardial adipose tissue.
BMI, 45 :t 3.4 kg/m2) who died of unnatural causes;
group 3-11 nonobese adults (weight, 84 :t 24 kg; BMI,
27 :t 3.9 kg/m2, currently classified as overweight) who
died of trauma. Epicardial fat was measured in the AV
grooves at the right and left lateral borders of the heart
and on the epicardial surface of the IVseptum, 2 cm distal
to the origin of the left anterior descending coronary
artery. The following is an epicardial fat index, calculated
as the mean of the 3 epicardial fat measurements taken in
each case: group 1, 11 :t 3.2 mm (:t SD); group 2, 11 :t 2.0
mm; group 3, 11 :t 2.0 mm. Thus, in this autopsy series,
mean EATthickness around the coronary arteries did not
differ over the BMI range of 27 to 57 kg/m2. Their results
cannot be directly compared with those of Schjebal's29
because their measurements were made in the AV
grooves and the anterior IV septum rather than the RV
free wall, as well as in patients selected according to body
weight and mode of death as opposed to a randomly
selected group of autopsy patients.
Radiology
In healthy people (n = 72) with BMI 22 to 47 kg/m2,
Iacobellis et al31used ECHO to measure epicardial fat and
found that the maximal thickness at any site over the free
wall of the RV varied between 1.8 and 16.5 mm. These
authors emphasized that they chose to measure epicardial
fat on the RV for 2 reasons: (i) this point is recognized as
the highest absolute epicardial fat layer thickness, and (ii)
their use of parasternal long- and short-axis views allow
the most accurate measurement of EAT on the RV, with
optimal cursor beam orientation in each view. Also using
Sacks and Fain 909
method Correlations Reference
31
ECHO EAT vs VAT': r = 0.84, P = .001
EAT vs SCAT': NS
32
ECHO EAT vs VATt: r = 0.80, P < .0001
EAT vs VAT/SCATt: r = 0.74, P = .0001
15
MRI EAT vs VAT: NS
PATt vs VAT: r = 0.66, P < .0006
PATt vs SCAT: NS
14
a PAT§ vs VAT: r = 0.81, P < .0001
PAT§ vs SCAT: not reported
17
a PAT§ vs VAT: r = 0.791, P < .001
PAT§ vs SCAT: r = 0.470, P < .001
PAT§ vs VAT: r = 0.692, P < .001
PAT§ vs SCAT: r = 0.410, P < .001
ECHO, Malvazos et al32reported mean EATvalues on the
free wall of the RV of 1.3 :t 0.2 mm (SD) in 15 healthy lean
(BMI, 22 :t 1.7 kg/m2) and 6.5 :t 0.8 mm in 27 healthy
obese (BMI,43 :t 4.8 kg/m2) women (P < .0001). Abbara
et al33point out that ECHO cannot give an adequate
window of all cardiac segments and is highly dependent
on acoustic windows, which are often inadequate for
subtle assessments in obese patients, resulting in an
insufficient examination.34 Abbara et al33measured EAT
using 16-slice scanner, multidetector computerized to-
mography (MDCn to assess CAD imaging in 59 adults
(BMI not reported) in a mapping study designed to
facilitate transepicardial arrhythmia ablation. The MDCT
has advantages of submillimeter collimation, high tem-
poral and spatial resolution, and 3-dimensional views of
the heart and its epicardial surface. The following are the
mean EATthickness at different sites in descending order
of magnitude: right AV groove, 14.8 mm; left AV groove,
12.7 mm; superior IV groove, 11.2 mm; inferior IV
groove, 9.2 mm; acute margin, 9.2 mm; anterior IV
groove, 7.7 mm; RVanterior free wall inferior, 6.8 mm; RV
anterior free waJI superior, 6.5 mm; RV superior wall,
5.6 mm; RV apex, 4.8 mm; LV apex, 2.8 mm; RV
diaphragmatic wall, 1.4 mm; and LVsuperior lateral wall,
1.0 mm. Mean EATthickness for all patients was 5.3 :t 1.6
mm (SD). Total EATcontent was on average 22% greater
for patients more than 65 years of age and 17% greater in
women in agreement with an autopsy report. 29There-
fore, in this cohort, the thickest part of EATwas in its
grooved and not, as some authors29"'1 suggest, in the
nongrooved segments that include the free wall of the RV.
910 Sacks and Fain
Magnetic resonance imaging (MRI) also has limitations for
EATdetermination in terms of its lower spatial resolution,
specifically in the through plane dimension (z-dimen-
sion).33 Nevertheless, it is considered the "gold standard"
for visceral fat measurement.35 The MRI and ECHO
measurements of EAT made over the free wall of the RV
correlate well (r = 0.91, P = .001).31
Correlations between EAT and PAT versus VAT and
SCAT determined radiologically are shown in Table 1. In
72 healthy adults with BMI 22 to 47 kg/m2, EAT
thickness determined by ECHO correlated significantly
with VAT measured by MRI (r = 0.84, P = .001) and
with waist circumference (r = 0.845, P = .01).31 The
association was less with BMI (P = .05), and there was
none with total fat mass (P = .1). Malavazos et al32
confirmed that EAT thickness by ECHO over the free
wall of the RV was significantly related to VAT (r = 0.8,
P < .0001). Sironi et al15 used MRI to measure EAT,VAT,
and PAT (mediastinal) fat in 13 hypertensive insulin-
resistant overweight men (BMI, 28 :!: 0.7 kg/m2 [SED
and 26 normotensive insulin-sensitive age. and BMI-
matched (27 :!: 0.5 kg/m2) controls. Unlike the ECHO
study discussed above,31 there was no correlation
between EAT and VAT areas in both groups. The
hypertensive group had significantly more PAT (45 :!: 5
vs 28 :!: 3 cm2; P = .005), but there was no difference
EAT area. There was a direct relationship between PAT
and VAT for the combined groups (n = 23; r = 0.66;
.0006). In a cohort of 69 patients with DM and 11 non-
DM siblings, PAT volume detemlined by computed
tomography (CT) correlated with VAT volume deter-
mined by CT (r = 0.81) compared to waist circumfer-
ence (r = 0.63) and BMI (r = 0.47),14 all being
significant (P < .0001).14 Epicardial adipose tissue
volume alone was not determined. Another striking
finding was the wide variability of pericardial fat
volumes ranging from 84 to 899 mL with a mean of
320 mL compared to mean VAT of 3046 mL. In a
Japanese study,17 PAT measured by CT in 181 nonobese
men with BMI 22.7 :!: 2.0 kg/m2 (SD) correlated with
both VAT (r = 0.791, P < .001) and with SCAT (r =
0.470, P < .001), as did PAT with VAT (r = 0.692, P <
.001) and with SCAT (r = 0.410, P < .001) in 64 obese
men with BMI of 27.6 :!: 2.39 kg/m2.17 In summary, in 2
CT studies, PAT correlated with VAT, but EAT per se
was not measured. In 2 studies, EAT determined by
ECHO correlated with VAT by MRI and CT. In one
study, EAT by MRI did not correlate with VAT, possibly
because of differences in the selection of subjects.
Pathophysiology of adipose tissue and
adipokines in obesity
The pathophysiology of adipokine expression and
secretion in VAT and SCAT needs to be reviewed to
provide a conceptual basis for understanding adipokine
American Hearl Jaurnai
June 2007
Figure 2
+ ---
-- ----
adipocyte h'JPerIrophy wth tItlycel1de
---------..
r--tVEGF. tleptin ITNFa. tMCP-l
If ! 1 +
! 1
ttissue vascularity trnacrophages
~l
1
tstromal capfllary angiogenesis monocyte recruitment into adipes.
:l::" +
t IL.10.t/t ;-r tI!:!!:::.tMCf'-1
11.-6.
adlpocylemuUn re_nee 1 adiponectln
1
t lipolysis. t
FFA release,
Pathophysiology and adipokine signaling in obese adipose tissue.
Adipocytes hypertrophy with triglyceride when energy intake
exceeds expenditure resulting in obesity. This triggers a cellular
and molecular inRammatory cascade with positive feedback loops
involving VEGF and MCP-]. The consequences are increased
adipose tissue vascularity along with enhanced accumulation of
macrophages and release of cytokines by the non-fat cells in
adipose tissue, local insulin resistance with accelerated lipolysis and
FFA release, and decreased adiponectin production and increased
leptin release byadipocytes. Plussign indicates stimulation; negative
in
sign, inhibition.
P <
pathophysiology. In contrast to the lack of studies
comparing EAT in healthy nonobese and obese humans,
the expression and secretion of adipokines in VAT and
SCAThave been well documented in biopsies taken from
lean healthy patients at elective intra.abdominal surgery
as compared with biopsies from obese healthy patients
undergoing bariatric procedures.36.39 In other studies,
omental VAT and SCATin obesity3,10.41or SCATfrom lean
and obese subjects42, 13or SCAT before and after weight
loss42,44have been determined. Collectively, the data
indicate that obese VAT and SCAT contain more macro-
phages, tumor necrosis factor-a (TNF-a), interleukin (1L)-
6, IL-8,IL-IO,resistin, monocyte chemoattractive protein-
I (MCP-I), plasminogen activator inhibitor-I, (PAl-I),
angiotensinogen (AGT), vascular endothelial growth
factor (VEGF), transforming growth factor I?> I, and less
adiponectin and leptin than lean VAT and SCAT.
Definitions of adipocytokine and adipokine vary.3-6We
define an adipokine as a hormone, cytokine, or chemo-
kine secreted from intact adipose tissue, which is
composed of a mixture of cells including adipocytes and
preadipocytes, macrophages, Iymphoctes, endothelial
cells, mast cells, basophils, and fibroblasts." A com-
monly used term for the nonadipocytes of adipose tissue
is the stromal-vascular matrix (SVM).44 IL-II~,IL-6, IL-8,
IL-IO, TNF-a as well as resistin and VEGF are released
primarily by the nonadipocytes (SVM), whereas the
American Hearl Journal
Vdllme 153, Number 6
chemokines, MCP-I, and macrophage migration
inhibitory factor,"5 as well as nerve growth factor and
serum amyloid Al and 2 proteins are produced to a
somewhat greater extent by human adipocytes.,,4 How-
ever, it should be emphasized that most inflammatory
cytokines released by obese human VAT and SCAT are
derived from the nonadipocytes, except for leptin and
adiponectin.3.44
The primary physiological role of adipocytes in
adipose tissue is as a depot in which to store fat as
energy when energy intake exceeds energy expenditure
and to release FFAs on demand." As an additional
function, when adipocytes in VAT and SCAT hypertro-
phy with triglyceride during obesity, they secrete TNF-a
and MCP-I, and the macrophage number increases
within the 2 fat depots that, as a result, transform into
inflammatory tissues46 (Figure 2). Circulating lympho-
cytes, and monocytes attracted by MCP-l secreted by
these expanding adipocytes,4:\,47 diapedese across the
endothelium of adipose tissue capillaries into the SVM.
Monocyte chemoattractive protein-I is essential for
promoting monocyte entry into the SVM as shown by
MCP-148and MCP-l receptor (CCR2)49 knockouts
that negate monocyte diapedesis. Recruited monocyte-
transformed macrophages in the SVM are termed
activated MI-polarized macrophages and express CCR2
(CCR2~.50 These macrophages secrete MCP-l to am-
plify monocyte recruitment, and together with activated
endothelium, also produce proinflammatory TNF-a, IL-
l~, IL-6, and IL-8,46.51which, with adipocyte-derived
TNF-a autocrine feedback, inhibit insulin signaling in
adipocytes via paracrine cross-talk. 52The result is
adipocyte insulin resistance and lipolysis of stored
triglyceride into FFAs.36.52.53In response, resident
CCR2~ alternatively activated M2-polarized macro-
phages increase the release of the anti-inflammatory
cytokine IL-IO to protect adipocytes from these inflam-
matory factors.5o Both TNF-a and IL-6 with its soluble
receptor inhibit adiponectin production. 54 The endo-
thelium produces VEGF,55and the adipocytes produce
leptin and VEGF, which stimulate angiogenesis to
increase adipose tissue vascularity parri passu with
adipocyte expansion when it occurs.46.56 Free fatty acids
released from VAT into the hepatic portal vein are
substrates for hepatic synthesis of atherogenic apopro-
tein B-containing VLDL-trigylceride particles that are
subsequently released into the peripheral circulation. 57
Pathophysiology of EAT
Epicardial fat in obesity
In their 1933 report on adiposity of the heart, Smith
and Willius58 performed autopsies on 136 obese patients
(mean 43% above ideal body weight; range, 13%-103%).
They noted that "in most instances, a definite relation-
ship between the excess of epicardial fat and the degree
Socksand Fain 911
of general obesity occurred." This observation was based
on increased heart weight and not on dissected
epicardial fat mass. As epicardial fat increases, it extends
over the anterior surface of the heart, more over the RV
than the LVand, lastly, over the LV midway between the
apex and base.13 The coronary arteries become encased
by or displaced in front of the enlarged epicardial fat
layer or lie between it and the myocardium. 58 The
amount of fat is variable and in extreme obesity can
cover the heart completely in fat 2 cm thick or more
("cor adipe plane tectum"). Fat also penetrates from the
subepicardial connective tissue into the connective
tissue lying between the muscle bundles and muscle
fibers, defined as adiposity of the heart.5!'!
Adiposity of the heart must be distinguished from
obesity-specific lipotoxic cardiomyopathy (LCMO),59 in
which excessive fat accumulates inside cardiac muscle
and causes LV remodeling and CMO, independent of
other causes of myocardial disease in obesity such as
hypertension and CAD.6oThe LCMO develops after
normal sites of fat storage in subcutaneous adipose
tissues, and VAT are filled to capacity in obesity and
release FFAs into blood. Excess circulating FFAs are
removed and converted into triglyceride by cardiomyo-
cytes as well as other cells in which small quantities of
fat are normally present, such as hepatocytes, skeletal
myocytes, and pancreatic islet ~-cells. The fat accumu-
lates intracellularly as droplets in the cytosol in these
"ectopic" sites, resulting, respectively, in myocardial
steatosis and a specific dilated CMO, nonalchoholic
steatohepatitis and cirrhosis, and DM.60At the molecular
level, the current hypothesis is that LCMO is not due to
triglyceride accumulation alone but is the consequence
of accumulating by-products of lipid metabolism such as
ceramide or other fatty acid derivatives that interfere
with intracellular signaling pathways through phospha-
tidylinositol 3-kinase and nuclear factor KB.59Ceramide
is a sphingosine signaling molecule that increases
inducible nitric oxide synthase activity and intracellular
nitric oxide leading to cardiomyocyte apoptosis.61
Putatively, FFAs released from hypertrophied adipocytes
in EAT could diffuse directly into the myocardium,
together with myocardial uptake of plasma FFAs,22
exacerbating myocardial steatosis, and lipotoxicity.
Structurally and functionally, the consequences of
intracardiac lipotoxicity and extracardiac adiposity in-
clude increased heart weight and mechanical pumping
effort,62 LV hypertrophy, LV diastolic dysfunction,
cardiac failure, electrocardiographic abnormalities, and
increased arrhythmogenicity.6"
The role of EAT in CAD
Obesity is an independent risk factor for CVD.63
Epicardial adipose tissue thickness, determined by ECHO
over the free wall of the RV, correlates with VAT (a CVD
risk factor per se), other correlates of CVD such as waist
912 Socks and Fain American Heart Jaurnai
June 2007

Table II. The relationship between EAT and CAD

Study (reference) Principle findings Limitations

Hypercholesterolemic white rabbits68
Human myocardial bridge69
Human anomalous coronary artery origin
from the sinus of Valsalva70.7l
Congenital generalized
Balloon overstretch injury of porcine
coronary arteries74
Human CAD/CABG surgery75
Human CAD/CABG surgery76
Human CAD/CABGsurge~
Human CAD/CABG surgery Heart
valve surgery78
Human autopsy coronary arterial
segments79
Human CADI7
Atherosclerotic lesions were absent in intra myocardial but Differences may have been due to
present in intraepicardial portions of the left anterior hemodynamic protective effects on
descending coronary artery surrounded by fat transendotheliallipid permeability
but a role for adipokines was plausible
Atherosclerotic intimal lesions were absent in the part Protection of intima may have been
of left anterior descending artery covered by myocardium due to hemodynamic forces during
while running through epicardial fat bridge contraction, but a role far
adipokines was feasible
Atherosclerotic intimal lesions were absent in the proximal Intima of intra-aortic coronary
coronary trunk lying in the subadventitial wall of aorta, segment might have been protected
despite distal CAD and multiple risk factors in some cases by hemodynamics during aortic
contraction in diastole
lipodystrophy72,73 Epicardial, visceral, and subcutaneous abdominal fat were It was unknown if the extent of surface
absent, yet CAD was found at autopsy lesions was worse or better than age- or
sex-matched controls
Macrophages and neutrophils occurred in and chemokines/ The pathophysiologic relevance to
cytokines were expressed from epicardial fat several millimeters lipoprotein-induced intima-media injury
from the site of adventitial injury was not clear
More mRNA for and secretion of MCP-], Il-] I>, Il-6, and Adipokines did not correlate with extent
TNF-a, and more chronic inRammatory cells were found in of CAD, risk factors, and BMI, and there
epicardial fat than leg subcutaneous fat were no data in controls without CAD
There was less adiponectin protein in epicardial fat than in No inRammatory cells were seen in CAD
controls with valvular heart disease without CAD and control epicardial fat, and no other
adipokines were measured
There was less adiponectin, Il-6, PAl-], and leptin mRNA, and Epicardial adipokines were compared to
more macrophage infiltration, resistin, and AGT mRNA in gluteal adipokines in separate patients
epicardial than gluteal subcutaneous fat rather than to epicardial adipokines in
controls without CAD
There was more TNF-a but similar adiponectin, MCP-] , Il-6, Epicardial and thoracic subcutaneous
resistin, and leptin mRNA in epicardial than thoracic fat data were pooled, and it was unclear
subcutaneous fat if valve patients had CAD
There was increased macrophage density in periadventitial BMIwas not reported
fat of coronaries with lipid cores compared to coronaries
with ~brocalci~c plaques or no atherosclerosis
Pericardial fat volume measured by a correlated most strongly Epicardial fat component of pericardial
with severity of angiographic atherosclerotic lesions than fat was not directly quantitated
other fat depots

circumference, diastolic blood pressure, plasma insulin,
fasting plasma glucose, high-density-lipoprotein-choles-
terol, low-density-lipoprotein-cholesterol, adiponectin,31
and with insulin resistance itself measured by the insulin
clamp technique.64 Based on these 2 studies by Iacobellis
et al,31 the authors have suggested that an increased
quantity of EAT in obesity may be a CVD risk predictor.
For confirmation, a larger, prospective, epidemiological
study, perhaps including direct measurements of peri-
coronary EAT thickness, should be done.
How could EAT mediate CAD? According to the
response-to-retention hypothesis, atherogenesis results
from the transendothelial passage (transcytosis) of
cholesterol-rich atherogenic Apo-B lipoproteins (very-
low-density-li poprotein, intermediate-density-lipopro-
tein, and low-density-lipoprotein) from plasma into
the intima, their retention in the subendothelial space
(a pivotal step), their oxidative modification, the
initiation and propagation of a chronic inflammatory
response in the intima, media, and adventitia leading to
plaque formation.65,66 If atherosclerosis is driven pri-
marily by luminal lipids and involves inflammation in
these 3 layers of the arterial wall, could inflammation in
periadventitial EAT also playa role in the pathogenesis
of CAD? 1,67
The principle findings and limitations of studies
examining the relationship between EAT and CAD are
presented in Table II.
The effect of the absence of EAT on CAD. In
hypercholesterolemic white rabbits, atherosclerotic
lesions are absent in intramyocardial but present in
intraepicardial portions of the left anterior descending
coronary artery surrounded by fat.68 In the human
"myocardial bridge," atherosclerotic intimal lesions are
not seen in the part of the left anterior descending
coronary artery covered by myocardium, which sepa-
rates the artery running through epicardial fat,69 In
anomalous origins of the coronary artery from the sinus
American Heart Journal
Volume 153, Number 6
of Valsalva,atherosclerotic intimal lesions are not
present in the proximal aberrant segment of the
coronary trunk as it runs through the subadventitial
layer of the aorta, even when multiple CAD risk factors
are present and despite evidence of distal atherosclero-
sis?O,71Although these anatomical "experiments of
natUre" suggesta plausible role for adipokines in
atherogenesis,an alternative explanation could be that
protective hemodynamic forces during cardiac or aortic
contraction reduce trans endothelial lipid permeability
into the intima. Autopsy proof of CAD despite total
absence of EAT, VAT, and SCAT in patients with
congenital generalized lipodystrophll.72 proves that
EAT is not necessary for atherosclerosis to develop and
progress but does not exclude a secondary role for it in
atherogenesis. The extent of surface lesions was not
quantified, which would have permitted a determination
of whether the severity of arterial disease in congenital
generalized lipodystrophy was worse or better than that
in age- and sex-matched normal controls considering
concomitant insulin-resistant DM and hyperlipidemia as
CVD risk factors.
Effects of experimental coronary artery
injury. In balloon overstretch injury of the media of
porcine coronary arteries, macrophages, neutrophils,
and adipokine expression were found in EAT several
millimeters from the site of adventitial injury compared
with few or none in controls/4 suggesting a signaling
system between media-adventitia and EAT. However,
the pathophysiological relevance of these findings to
lipoprotein-induced intimal injury is unclear.
Epicardial adipose tissue adipokines and
histopathology at CABG. Tumor necrosis factor-a,
MCP-I, 11.-1~\ and 11.-6mRNA expression and secretion,
and chronic inflammatory cell infIltration with macro-
phages, lymphocytes, and basophils were increased in
EAT versus leg subcutaneous fat from obese patients
(BMI, 31 1: 1 kg/m2 [SD]; n = 42) with multiple
cardiovascular risk factors undergoing coronary artery
bypass graft for critically stenotic CAD.75Epicardial
adipose tissue inflammation did not seem to result from
atherogenic inflammation in the underlying plaques
(mediated by "inside-to-outside" signaling) and was not
related to circulating cytokines implying that obese EAT
per se might be proinflammatory. Notably, adipokine
expression and secretion in a control group of patients
without multiple risk factors or with noncritical CAD
undergoing open-heart surgery for other indications were
not studied, In addition, adipokines did not correlate with
the extent of CAD, risk factors, and BMI. Adiponectin
improves insulin sensitivity and has anti-inflammatory
and antiatherogenic actions so that low adiponectin
levels within the arterial wall, and in the circulation in
metabolic syndrome and DM, may result in the loss of its
potential vasoprotective effects.8o.H2Adiponectin protein
levels in EATwere lower from nonobese patients (BMI,
Sacks and Fain 913
Figure 3
AOIPOKINES
Adlponectln
Leptin
TN Fa
IL-113
MCP-1
PAI.1
@ IL.a
IL-S
Reslstln
Anglotensinogen
VEGF
Hypothetical mechanisms whereby epicardial adipokines might play
a role in coronary atherogenesis. Of the adipokines identified in
human EAT (inset), adiponectin and leptin are produced exclusively
by adipocytes. The other adipokines are expressed in varying
amounts by both adipocytes and stromal preadipocytes, macro-
phages, lymphocytes, fibroblasts, and endothelium. Pathway 1:
Paracrine signaling. Adipokines secreted from adipocytes and
stromal-vascular cells in EAT overlying the lipid core of athero-
sclerotic plaques diffuse in interstitial nuid across the adventitia,
media, and intima and interact respectively with vasa vasora,
vascular smooth muscle cells, endothelium, and cellular components
of the plaque. Paracrine signaling may also occur between
adipokines and FFA diffusing from epicardial fat into the underlying
myocardium (not shown). Pathway 2: Vasocrine signaling. Adipo-
kines secreted by epicardial adipocytes and stromal-vascular cells
closely apposed to adventitial vasa vasorum traverse the vessel into
its lumen and are transported downstream to react with cells in the
media and the intima around plaques. In this model, macrophages
and lymphocytes can migrate alongside the vasa vasorum through
breeches in the media.90
27.4 1: 3.5 kg/m2 [SD]) with CAD compared with controls
(BMI, 25.7 1: 3,2 kg/m2) without CAD.76In this stUdy,
other adipokines were not measured, and no inflamma-
tory cells were seen in CAD and control EAT.Adiponectin
mRNA in EAT from 46 nonobese patients (BMI, 27 1:
3.3 kg/m2 [SEM])with CAD was lower than adiponectin
mR.J'IAin omental, SCAT, and gluteal fat from 30 non-
obese patients (BMI, 25.7 :t 4.7 kg/m2) without CAD77
confirming this trend, but the data can be questioned
because of inappropriate controls. In addition, less IL-6,
PAl-I, and more resistin, AGT, and macrophage CD45
mRNA was found in EATversus gluteal fat. By contrast, in
a smaller group of 15 patients (10 CABG and 5 valve
replacements; BMI, 26.6 1: 1.2 kg/m2 [SEM]), TNF-a
mRt'lA expression in EATwas increased compared
to subcutaneous thoracic fat, whereas there were no
914 Socks and Fain
differences in adiponectin, IL-6, MCP-I, resistin, and
leptin mlL"IIAin the 2 fat depots.78 It was not clear
whether the valve patients had CAD or not.
Human autopsy coronary arterial
segments. CD68 macrophages were more numerous
and densely packed in the periadventitial fat of athero-
sclerotic arteries with lipid cores than in that of
fibrocalcific or nonatherosclerotic arteries.79 Body mass
index was not reported. This suggested that macrophages
could enter a plaque through the adventitia or periad-
ventitial fat even in the early stages of atherosclerosis.
Radiology. Pericardial fat area, including EAT, mea-
sured by thoracic CT correlated significantly with the
extent of CAD measured angiographically in both lean
(BMI, -23) and overweight (BMI, -28) nondiabetic
Japanese men.17 However, it is not clear to what extent
EAT area per se correlated with CAD in this study.
Adipokine paracrine and vasocrine signaling
If EATdoes contribute to atherogenesis, how might this
come about? Figure 3 depicts hypothetical mechanisms
whereby adipokines generated in EAT around athero-
sclerotic coronaries may access plaques in the intima. IL-
l J»experimentally applied to the arterial adventitia can
cause inflammatory changes by diffusion into the intimal
layer.83 Thus, it is plausible that paracrine release of
cytokines from periadventitial EAT could traverse the
coronary wall by diffusion from "outside-to-inside" and
interact with cells in each of its layers. Likewise, it has
been suggested that adipokines released by macrophages
and lymphocytes aggregating at the adventitia-fat inter-
face of an atherosclerotic aortic aneurysm could diffuse
into the intima-media.84 On the other hand, during
atherogenesis, cellular proliferation and plaque formation
can increase the arterial wall thickness to 3 to 4 mm
compared to 0.55 to 1.0 mm normallyZ5 so that adipokine
diffusion might become less important than vascular
access. In this respect, adipokines and FFAs might be
released from epicardial tissue directly into vasa vasorum
and be transported downstream into the arterial wall
(Figure 3). This process termed "vasocrine signaling" is
derived from studies by Judkin et al85on adventitial fat
around arterioles supplying the cremaster muscle in the
obese rat and may be applicable to second-order vasa
vasora of similar arteriolar caliber.86 Vasa vasora arise
from bifurcation segments of the epicardial coronary
arteries within the adventitia and divide into first-order
parallel and second-order circumferential branches that
penetrate the coronary wall to supply oxygen and
nutrients to its outer tWo thirds, while the inner third is
supplied by diffusion from the lumen.87
In diabetic animal models and humans with CAD,
adventitial inflammation and vasa vasorum neogenesis
are responsible for neovascularization of both media and
intimal plaques.88.89 The inflammatory response in the
adventitia is characterized by accumulation of macro-
American Hearl Journal
June 2007
phages and B-Iymphocytes. 1.90As the vasa vasora create
breeches in the media wall, they become surrounded
mainly by foci of T lymphocytes and perivascular
macrophages.90 These breeches may possibly be medi-
ated by T-helper cell-driven immune responses via
interferon--y, which inhibits vascular smooth muscle
proliferation contributing to medial disruption. Vasa
vasorum neogenesis is mediated by VEGF secreted by
activated T lymphocytes in the intima-media,90 but
conceivably also by epicardial adipocytes, because
adipocytes harvested from human donor EAT secrete
VEGF and induce angiogenesis of coronary artery
endothelial cells in vitro. 56 Virmani et al90 have empha-
sized the critical role of vasa vasorum neovascularization
in plaque hemorrhage, stability, and rupture.
Areas of future research for EAT
Current evidence implies that EATis a contributor to
the progression of CADrather than an "innocent
bystander" (an epiphenomen) 1 or an associated marker.
More is needed to support this hypothesis.
Histopathology. Macrophages in plaques and adi-
pose tissue are heterogeneous,50,91 and subsets pos-
sessing distinct surface antigens or receptors specific
for EAT and intima macrophages92.93 may be identified
and used to track the intramural movement of these
cells bidirectionally. In addition, macrophage density
should be measured in autopsy specimens from obese
and lean patients without CAD. Superparamagnetic iron
oxide nanoparticles, phagacytosed by macrophages,
might be used as a contrast agent to enhance the
spatial resolution of MRI for delineating active plaques
in coronary arteries and inflammatory activity in EAT.8z
Animal studies. Experiments should be performed
in animals with normally discernible EAT such as
rabbits,66 pigs,86 or monkeys94 with high fat- or
cholesterol-induced hyperlipidemia and atherosclerosis,
rather than rodents that nortllally have little or no EAT,18
to examine the relationship of coronary atherogenesis to
adipokine expression in EAT. Because adventitial vasa
vasorum angiogenesis precedes vascular lesion forma-
tion,86 the initial molecular signals responsible for
adventitial angiogenesis could arise, at least partly, from
secretion of VEGF and leptin by hypertrophied adipo-
cytes or other cell components present in EAT(Figure 2).
Clinical studies. These should examine whether (i)
EAT is a biomarker of CAD, peripheral arterial disease, or
cerebrovascular disease, and (ii) whether EAT offers
incremental value over traditional CVD risk factors as
predictors of cardiovascular outcomes. Because piogli-
tazone suppresses SCAT macrophage number,95 a study
comparing the effect of a thiozolidenedione on adipo-
kine expression and macrophage content in EAT in
patients with DM at CABG would be useful. Biopsies of
EAT obtained during CABG should be compared with
EAT biopsies from weight-matched controls without
American Heart Journal
Volume 153, Number 6
CAD. Lastly, the effect of caloric restriction, weight loss,
or pharmacological agents on EAT could be determined
by radiological methods such as MDCT,
We thank Dr Bruce Webber for providing and
reviewing the histology slide, Doctor George Cowan
gave helpful comments in reviewing the manuscript.
References
1, Chaldkov GN, Fiore M, Ghenev PI, et 01. Atherosclerotic lesions:
possible interactive involvement of intima, adventitia and associated
adipose tissue, Int Med 2000;7:43-4,
2, Wajchenberg BL Subcutaneous and visceral adipose tissue. Their
relation to the metabolic syndrome. Endocr Rev 2000;21 :697 -738.
3. Fain IN, Madan AK, Hiler ML, et 01. Comparison of the release of
adipokines by adipose tissue, adipose tissue matrix and adipocytes
from visceral and subcutaneous abdominal adipose tissues of obese
humans. Endocrinology 2004; 145:2273 - 82.
4. Kershaw EE, Flier JS. Adipose tissue as on endocrine organ. J Clin
Endocrinol Metab 2004;89:2548-56.
5. Trayhurn P. Endocrine and signalling role of adipose tissue; new
perspectives on fat. Acta Physiol Scand 2005; 184:285 -93.
6. Scherer P. Adipose tissue. From lipid storage comportment to
endocrine organ. Diabetes 2006;55: 1537 -45.
7. Weiss R, Dziura J, Burgert TS, et 01. Obesity and the metabolic
syndrome in children and adolescents. N Engl J Med
2004;350:2362 -74.
8. Ford ES, Giles WH, Dietz WHo Prevalence of the metabolic
syndrome among US adults. JAMA 2002;287:356-3598.
9. Lakka HM, Laakonsen DE, Lakka T, et 01. The metabolic syndrome
and total and cardiovascular disease mortality in middle-aged men.
JAMA 2002;288:2209-716.
10. Eckel RH, Baruch WF, Ershow AG. Report of the Notional Heart,
Lung and Blood Institute-Notional Institute of Diabetes and Digestive
Disease and Kidney Diseases working group on the pathophysiol-
ogy of obesity-associated cardiovascular disease. Circulation
2002;105:2923-8.
11. Ho E, Shimada Y. Formation of the epicardium studied with the
scanning electron microscope. Dev Bioi 1978;66:579 - 85.
12. Wililiams PL The anatomicol basis of medicine and surgery. Gray's
Anatomy. 38th ed. Philadelphia, PA: Churchill Livingstone; 1995.
p.1493.
13. lacobellis G, Corradi D, Sharma AM. Epicardial adipose tissue:
anatomic, biomolecular and clinical relationship with the heart.
Nature clinical practice. Cardiovas Med 2005;2:536-43.
14. Wheeler GL, Shi R, Beck SR, et al. Pericardial and visceral adipose
tissue measured volumetrically with computed tomography are highly
associated in type 2 diabetic families. Invest RadioI2005;40:97 -101.
15. Sironi AM, Gastaldelli A, Mari A, et al. Visceral fat in hypertension:
inRuence of insulin resistance and b-cell function. Hypertension
2004;44: 127 - 33.
16. Moore KL,Persaud TVN. The developing human. Clinically oriented
embryology. 7th ed. Philadelphia, PA: Saunders; 2003. p. 189.
17. Taguchi R, Takasu J, Itani I, et al. Pericardial fat accumulation in men
as a risk factor for coronary artery disease. Atherosclerosis
2001 ;157:203- 9.
18. Marchington JM, Mattocks CA, Pond CM. Adipose tissue in the
mammalian heart and pericardium; structure, foetal development
and biochemical properties. Comp Biochem Physiol
1989;94B:225 -32.
Sacks and Fain 915
19. Marchington JM, Pond CM. Site-specific properties of pericardial
and epicardial adipose tissue: the effects of insulin and high-fat
feeding on lipogenesis and the incorporation of fatty acids on vitro.
Int JObes 1990;14:1013-22.
20. McCance KL,Huether SE. Chambers of the heart. Pathophysiology.
The biologic basis for disease in adults and children. 4th ed. St.
Louis, MO: Mosby; 2002. p. 932.
21. Wisneski JA, Gertz EW, Nease RA, etal. Myocardial metabolism of
free fatty acids. J Clin Invest 1987;79:359 -66.
22. Nelson RH, Prasad A, Lerman A, et al. Myocardial uptake of
circulating triglycerides in non-diabetic patients with heart disease.
Diabetes 2007;56:527 - 30.
23. Bober SR, Deng W, Rodriguez J, et 01. Vasoactive prostanoids are
generated from arachidonic acid by COX-l and COX-2 in the
mouse. Am J Physiol Heart Circ Physiol 2005;289:HI476-87.
24. Gollasch M, Dubrovska G. Paracrine role for periadventitial
adipose tissue in the regulation of arterial tone. Trends Pharmacol
Sci 2004;25:647 - 53.
25. Fayad ZA, Fuster V, Fallon JT, et al. Non-invasive in vivo human
coronary lumen and wall imaging using block-blood magnetic
resonance imaging. Circulation 2000;102:506-10.
26. Fesus G, Dubrovska G, Essin K, et al. Adiponectin is a novel potent
humoral vasodilator. Acta Physiol 2006; 186(Suppl 650):OW02.
27. Corradi D, Maestri R, Callegari S, et al. The ventricular
epicardial fat is related to the myocardial moss in normal,
ischemic and hypertrophic hearts. Cardiovasc Pathol 2004;
13:313-6.
28. lacobellis G, Ribaudo MC, Zappaterreno A, et al. Relation between
epicardial adipose tissue and left ventricular mass. Am J Cardiol
2004;94: 1084 -7.
29. Schejbal V. Epicardial fat on the right ventricle-morphology,
morphometry and functional significance. Pneumologie
1989;43:490- 9.
30. DuHou J, Virmani R, Rabin I, et al. Sudden death as a result of heart
disease in morbid obesity. Am Heart J 1995; 130:306 - 13.
31. lacobellis G, Ribando MC, Assai F, et al. Epicardial adipose tissue is
related to anthropametric and clinical parameters of metabolic
syndrome: a new indicator of cardiovascular risk. J Clin Endocrinol
Metab 2003;88:5163-8.
32. Malavazos AE, Ermetici F, Coman C, et al. InHuence of epicardial
adipose tissue and adipocytokine levels on cardiac abnormalities in
visceral obesity. Int J Cardiol 2006 [in press].
33. Abbara S, Desai JC, Ricardo CC, et al. Mapping epicardial fat with
multidetector computed tomography to facilitate percutaneous
transepicardial arrhythmia ablation. Eur J Radiol 2005;57:417 - 22.
34. Kessels K, Cramer MJ, Veldhuis B. Epicardial adipose tissue imaged
by magnetic resonance imaging: an important risk marker of
cardiovascular disease. Heart 2006;92:262.
35. lacobellis G. Imaging of visceral adipose tissue: on emerging
diagnostic tool and therapeutic target. Curr Drug Targets Cardio-
vase Hematol Dis 2005;5:345 - 53.
36. Hotamisligil GS, Arner P, Caro JF, et al. Increased adipose tissue
expression of tumor necrosis factor-a in human obesity and insulin
resistance. J Clin Invest 1995;95:2409 -15.
37. Lefebre AM, Laville M, Vega N, et al. Depot-specific differences in
adipose tissue gene expression in lean and obese subjects. Diabetes
1998;47:98 -1 03.
38. Von Eyben FE, Kroustrup JP, Larsen JF, et al. Comparison of gene
expression in intra-abdominal and subcutaneous fat. A study of men
with morbid obesity and non-obese men using microarray and
proteomics. Ann N Y Acad Sci 2004;1030:508-36.
916 Sacks and Fain American Heart Jaurnai
June 2007

39. D'adamo M, Consoli C, Guglielmi V, et 01. Role of macrophage 58. Smith HL, Willi us FA. Adiposity of the heart: A clinical study of one
infiltration in inflammatory changes of obese subjects' adipose hundred and thirty six obese patients. Ann Intern Med
tissue. Diabetes 2006;55(Suppll):371-0R [abstrl. 1933;52:911 - 31.
40. Fried SK, Bunkin DA, Greenberg AS. Omental and subcutaneous 59. Schaffer JE. Lipotoxicity: when tissues overeat. Curr Opin lipidol
adipose tissues of obese subjects release interleukin-6: depot 2003;14:281-7.
difference and regulation by glucocorticoid. J Clin Endocrinol Metab 60. McGavrock JM, Victor RG, Unger RH, et 01. Adiposity of the heart,
1998;83:847 - 50. revisited. Ann Intern Med 2006; 144:517 - 24.
41. Cancello R, Tordjman J, Poitou C, et 01. Adipose tissue macro- 61. Zhou YT, Grayburn P, Karim A, et 01. Lipotoxic heart disease in
phages(ATM): omental white adipose tissue(WAT) is more infiltrated obese rats: implications for human obesity. Proc Natl Acad Sci USA
than subcutaneous WAT in human obesity. Diabetes 2006; 2000;97:1784-9.
55(Supp11 ):369-0R [abstrl. 62. lacobellis G, Sharma AM. Ann Intern Med 2006;145:554 [Letter].
42. Kern PA, Saghizadeh M, Ong JM, et 01. The expression of tumor 63. Poirer P, Giles TD, Bray GA, et 01. Obesity and cardiovascular
necrosis factor in human adipose tissue. Regulation by obesity, disease. Pathophysiology, evaluation, and effect of weight loss.
weight loss and relationship to lipoprotein lipase. J Clin Invest ArteriosclerThromb Vasc Bioi 2006;26:968 -76.
1995;95:2111-9. 64. lacobellis G, Leonetti F. Epicardial adipose tissue and insulin
43. Weisberg PS, McCann D, Desai M, et 01. Obesity is associated with resistance in obese subjects. J Clin Endocrinol Metab
macrophage accumulation in adipose tissue. J Clin Invest 2005;90:6300 - 2.
2003;112: 1796-808. 65. Williams KJ, Tabas I. The response-to-retention hypothesis of
44. Fain IN. Release of interleukins and other inflammatory cytokines by atherogenesis reinfarced. Curr Opin lipidoI1998;9:471 - 4.
human adipose tissue is enhanced in obesity and primarily due to 66. Williams KJ, Tabas I. lipoprotein retention-and clues for atheroma
the non-fat cells. In: Litwack G, editor. Interleukins: In Vitamins and regression. Arterioscler Thromb Vasc Bioi 2005;25: 1536 -40.
Hormones. Vitam Horm 2006; 74:443-77. 67. Chaldkov GN, Stonkulov IS, Aloe L. Subepicardial fat in human
45. Skuple PT, Harder C, Kraft I, et 01. Production and release of coronary atherosclerosis: another neglected phenomenon. Athero-
macrophage migration inhibitory factor from human adipocytes. sclerosis 2001; 154:237 - 8.
Endocrinology 2005; 146: 1006 - 11. 68. Ishikawa Y, Ishii T, Asuwa N, et 01. Absence of atherosclerosis
46. Wellen KE, Hotamisligil GS. Obesity-induced inflammatory changes evolution in the coronary arterial segments covered by myocar-
in adipose tissue. J Clin Invest 2003;112:1785-8. dial tissue in cholesterol-fed rabbits. Virchows Arch 1997;430:
47. Xu H, Barnes GT, Yang Q, et 01. Chronic inflammation in fat plays 163-71.
a crucial role in the development of obesity-related insulin 69. Ishii T, Asuwa N, Masuda S, et 01. The effects of a myocardial
resistance. J Clin Invest 2003; 112: 1821 - 30. bridge on coronary atherosclerosis and ischemia. J Pathol
48. Kanda H, Tateya S, Tamori Y, et 01. MCP-l contributes to 1998;185:4-9.
macrophage infiltration into adipose tissue, insulin resistance and 70. Angelini P, Velasco JA, Ott D, et 01. Anomalous coronary artery
hepatic steatosis in obesity. J Clin Invest 2006; 116: 1494 -505. arising from the opposite sinus: descriptive features and patho-
49. Weisberg SP, Hunter D, Huber H, et 01. CCR2 modulates physiologic mechanisms, as documented by intravascular ultra-
inflammatory and metabolic effects of high fat feeding. J Clin Invest sound. J Invasive Cardiol 2003; 15:507 - 14.
2006;116:115-24. 71. LitovskyS, Henderson J, TallajJA, et 01.Acute take-offof the right
50. Lumeng CN, Bodzin JL, Saltiel AR. Obesity induces a phenotypic coronary artery with long intra-aortic wall course presenting as
switch in adipose tissue macrophage polarization. J Clin Invest chronic coronary ostium occlusion in a patient with end-stage heart
2007:175-84. failure. J Heart Lung Transplant 2006;25:740 - 1.
51. Bruun JM, lihn AS, Madan AK, et 01. Higher production of IL-8 in 72. Case records of the Massachusetts General Hospital. N Engl J Med
visceral vs. subcutaneous adipose tissue. Implication of nonadipose 1975;292:35 -41.
cells in adipose tissue. Am J Physiol Endocrinol Metab 73. Chandalia M, Garg A, Vuitch F, et 01. Postmortem findings in
2004;286:E8-EI3. congenital genera Iised lipodystrophy. J Clin Endocrinol Metob
52. Suganami T, Nishida I, Ogawa Y. A paracrine loop between 1995;80:3077 - 81.
adipocytes and macrophages aggravates inflammatory changes. 74. Okamoto E, Couse T, De Leon H, et 01. Perivascular inflammation
Role of free fatty acids and tumor-necrosis factor-",. Arterioscler after balloon angioplasty of porcine coronary arteries. Circulation
Thromb Vasc Bioi 2005;25:2062 - 8. 2001;104:2228 -35.
53. BruunJM, Verdich C, Toubro S, et 01.Associationbetween 75. Mazurek T, Zhang L, Zalewski A, et 01. Human epicardial adipose
measures of insulin sensitivity and circulating levels of interleukin-8, tissue is a source of inflammatory mediators. Circulation
interleukin-6 and tumor necrosis factor-a. Effect of weight loss in 2003;108:2460-6.
obese men. Eur J Endocrinol 2003; 148:535 -42. 76. lacobellis G, PistilliD, Gucciardo M, et 01. Adiponectin expression in
54. Bruun JM, lihu AS, Verdich C, et 01. Regulation of adipose tissue- human epicardial adipose tissue is lower in patients with coronary
derived cytokines: in vivo and in vitro investigations in humans. Am J artery disease. Cytokine 2005;29:251 -5.
Physiol Endocrinol Metab 2003;285:E527-33. 77. Baker AR, do Silva NF, Quinn AL, et 01.Human epicardial adipose
55. Yoon Y, Losordo DW. All in the family. YEGF-B joins the ranks of tissueexpressesa pathogenic profile of adipocytokines in patients
proangiogenic cytokines. Grc Res 2004;93:87 - 90. with cardiovascular disease. Cardiovasc Diabetol2006;5: 1-7.
56. Stoll LL,Romig-Martin SA, Harrelson AL, et 01. Isolation and 78. Kremen J, Dolinkova M, Krajickova D, et 01. Increased subcutaneous
characterization of human epicardial adipocytes: potential role in and epicardial adipose tissue production of proinflammatory
vascular inflammation. Exp Bioi 2006;20: 1074. cytokines in cardiac surgery patients: possible role in post-operative
57. Ginsberg HN. Insulin resistance and cardiovascular disease. J Clin insulinresistance.J Clin EndocrinolMetab 2006 [published ahead
Invest 2000; 106:453 - 8. of print].
American Hearl Journal
Sacks and Fain 917
Volume 153, Number 6

79, Vela D, Buja M, Madjid M, et aL The role of periadventitial fat in 88. Moreno PR, Fuster V. New aspects in the pathogenesis of diabetic
atherosclerosis. An adipose subsetwith potential diagnostic and atherothrombosis, J Am Coli Cardiol 2004;44:2293 - 399.
therapeutic implications. Arch Pathol Lab Med 2006; 130:108 - 15, 89, Hayden MR, Tyagi sc. Vasa vasorum in plaque angiogenesis,
80, Matsuzawa Y, Funahashi T, Kihara S, et aL Adiponectin and metabolic syndrome, type 2 diabetes mellitus and atheroscleropathy:
the metabolic syndrome. Arterioscler Thromb Vasc Bioi a malignant transformation. Cardiovasc Diabetol 2004;3: 1 - 16.
2004;24:29-33, 90. Virmani R, Kolodgie FD, Burke AP, et al. Atherosclerotic plaque
81, Goldstein BJ, Scalia R, Adiponectin: a novel adipokine linking progression and vulnerability to rupture, Angiogenesis as a source
adipocytes and vascular function. J Clin Endocrinol Metab of intraplaque hemorrhage. Arterioscler Thromb Vasc Bioi
2004;89:2563-8, 2005;25:2054-61.
82, PisconT, Girman CJ, Hotamisligil GS, et aL Plasma adiponectin 91, Gordon S, Macrophage heterogeneity and tissue lipids. J Clin Invest
levels and risk of acute myocardial infarction in men. JAMA 2007;117:89-93.
2004;291: 1730-7, 92, Tacke F, Alvarez D, Kaplan TJ, et aL Monocyte subsets employ
83, Shimokawa H, Ito A, FukumotoY, et aL Chronic treatment with IL- CCR2, CCR5, and CX3CR 1 to accumulate within atherosclerotic
l beta inducescoronary intimal lesionsand vasosposticresponsesin plaques, J Clin Invest 2007;117: 185-94,
pigs in vivo, J Clin Invest 1996;97:769-76, 93. Swirski FK, Libby P, Aikawa E, et aL Ly-6Chi monocytes dominate
84. Henrichot E, Juge-Aubry CE, Pernin A, et aL Production of hypercholesterolemia-associated monocytosis and give rise to
chemokines by perivascular adipose tissue.Arterioscler Thromb macrophages in atheromata. J Clin Invest 2007;117:195-205.
Vasc Bioi 2005;25:2594-9, 94. Rudel LL, Parks JS, Sawyer JK. Compared with dietary monounsat-
85. Judkin JS, Eringa E, Stehouwer CDA "Vasocrine signalling" from urated and saturated fat, polyunsaturated fat protects African green
perivascular fat: a mechanism linking insulin resistanceto vascular monkeys from coronary artery atherosclerosis. Arterioscler Thromb
disease, Lancet2005;365: 1817 - 20. Vasc Bioi 1995;15:2101-10,
86, Kwon HM, Sangiori G, Ritman EL,et ai, Enhanced coronary vasa 95, Di Gregorio GB, Yao-Borengasser A, Rasouli N, et aL Expression of
vasorum neovascularisation in experimental hypercholesterolemia.J CD68 and macrophage chemoaltractant protein-l gene expression
Clin Invest 1998;101:1551-6, in human adipose and muscle tissues. Association with cytokine
87. Moreno PR, PurushothamanKR, Sirol M, et aL Neovascularisation expression, insulin resistance, and reduction by pioglitazone,
in human atherosclerosis.Circulation 2006; 113:2245 -52, Diabetes 2005;54:2305 -13,